Summary
%is paper has its departure point in ten interviews conducted by the famous journalist Larry 
King on the CNN. It sets out to demonstrate that fillers, words and phrases that may carry a 
negligible if not void semantic load, may, nevertheless, play an important functional role in 
communication, as discourse markers and interactional signals. 
Key words: filler, discourse marker, interactional signal, symbol function, symptom function
 Povzetek
Izhodišče članka je deset intervjujev znanega novinarja družbe CNN Larryja Kinga. Namen 
članka je dokazati, da mašila, beside in besedne zveze, ki imajo zanemarljav semantični pomen 
oziroma ga sploh nimajo, lahko v sporazumevanju igrajo pomembno vlogo kot diskurzni 
označevalci ali sporazumevalni sigali.
Ključne besede: mašilo, diskurzni označevalec, sporazumevalni signal, simbolična funkcija, 
simptomatična funkcija
 DOI: 10.4312/elope.7.2.45-56

Dialogues are types of spoken interaction, here-and-now social activities that, according to 
Stenström (1994, 3), are “governed by two main principles: speakers take turns and they 
cooperate”. %is, however, does not mean that the listener always waits for the speaker to 
finish before taking over; nor does it mean that the participants in the dialogue are always in 
full agreement with each other, that they never contradict each other, that objections to what 
is being said are never made or that surprising changes of topic never occur. Since speech is 
delivered spontaneously, hesitations, false starts, repetitions or incomplete utterances are fairly 
natural in a dialogue. And last, but not least, emotions are always engaged in a verbal exchange 
between two or more people.  
Most of these characteristics of dialogues are embodied in the language used by the participants 
in this kind of oral communication. Sounds, words, phrases, or even clauses that are “relatively 
syntax-independent, do not have a particular grammatical function, do not change the meaning 
of the utterances and have a somewhat empty meaning” themselves (Lee 2004, 117) are illustrative 
of such language. 
%ey have been paid little attention by linguists on the very grounds that their semantic 
contribution to the message is, most of the time, negligible, if not nil. It may be due to the reduced 
interest they have stirred that the terminology referring to them has been inconsistent. %us, some 
scholars, such as Lee (2004), include all these linguistic items in the category of fillers and point 
out that some of them function as markers of discourse development); others, such as Svartvik 
(1980) and Stubbs (1983), while calling them by the same name, analyze only few of the words in 
this class (basically, “oh” and “well”, referring to them also as adverbs, interjections, particles). Still 
other linguists classify them differently. Schiffrin (1987), for example, abandons the label fillers in 
favour of discourse markers, which she defines operationally as “sequentially dependent elements 
which bracket units of talk”.  On the other hand, for Stenström (1994, 223), fillers are only those 
“lexically empty items with uncertain discourse functions, except to fill a conversational gap”. 
Words and phrases without a referential meaning that are “used to organize and hold the turn 
and to mark boundaries in the discourse” (Stenström 1994, 63) are considered discourse markers. 
%ey “help the speaker organize the discourse. %ey serve to start a conversation; they serve to 
introduce and mark the end of a topic; and they signal the end of a conversation”. Non-referential 
words and phrases that “are used to start, carry on and terminate the conversation” are called 
by her interactional signals (Stenström 1994, 61), a definition which does not seem to be very 
helpful when it comes to making a clear difference between what she calls discourse markers and 
interactional signals. %e fact that the latter “appeal for feedback and give feedback, they respond, 
they involve the listener in the conversation, and so on; in other words, they play a crucial role for 
a smooth interaction” (Stenström 1994, 61),  may shed some light on the matter.
In the present paper, the term filler is used for any word or phrase that has no referential meaning, 
makes no semantic contribution to the content of the message transmitted, and is syntactically 

detachable from the sentence in which it occurs, in other words, a word or phrase whose absence 
would leave the propositional content of the verbal exchange intact. 
On the basis of an analysis of fillers carried out  on the basis of ten interviews conducted by the 
well-known journalist Larry King on CNN (see Corpus), this paper will, however, highlight also 
a number of specific cases in which one has to admit to the fillers’ playing some role (other than 
semantic in communication), in spite of their being considered useless signs of bad language, 
“words that jump out of the mouth while we are figuring out what clever things to say next” 
(Andersson and T rudgill 1992, 95). It is undeniable that fillers have no other role than to fill a 
gap in conversation in examples such as (1) to (3):
(1) RG: Well, to tell you the truth, I think, even the first time I was hit, it was a turn.
(2)  RG: Well, I was so numb. I mean, certainly, I talked – Barbara knew what was going on. 
(3) SB: …it was really comfortable for people like myself, who just sort of like to laugh…
However, my paper reveals several communicative situations in which fillers in the category of 
verbs (come on, look, listen), nouns (man), adverbs (yes, right, well, just, now, OK, clearly, totally), 
interjections (oh), phrases, of which some are exclamatory (sort of, all right, for sure, in some ways, 
for god’s sake, my gosh, my goodness, oh, dear, oh, boy), tag questions, non-finite clauses (to tell you 
the truth) and complete or incomplete finite clauses (I mean, I think, the point is, you know, you 
see, let me see, I’ll tell you what, I’m telling you) may have some unquestionable functional role.  
 In this paper, the general functions of fillers are theoretically seen from three points of view 
that add to each other to articulate a coherent approach to what the role in communication 
of these linguistic devices may be – Bühler’s (1990) Organon model of language, Schourup’s 
(1985) worlds of conversation and Schiffrin’s (1987) perspective on the flow of information in 
interactive discourse and how this is marked.  
If looked at from the perspective of Bühler’s (1990) Organon model of language, fillers have 
no symbolic, representation function, i.e. they have no referents in the surrounding reality, and 
thus they cannot serve the purpose of “one’s person communicating with another about things” 
(Bühler 1990, 35). However, they have both a symptom or expressive function, i.e. they provide 
information about the speaker’s attitude towards his/her words or his/her interlocutor’s words, 
show the speaker’s attitude towards the interlocutor or replace whole statements that should have 
described emotional reactions, and a signal or appeal function, i.e. they are directed towards the 
interlocutor and meant to trigger a certain (linguistic) behaviour or attitude on his/her part.
Partly similarly, if seen from the point of view of Schourup’s (1985) description of the worlds of 
conversation, fillers may be considered linguistic elements that provide hints at the relationship 
between the private worlds (of knowledge, beliefs, feelings, hopes, etc.) of the participants in 
conversation and the shared world they are building while communicating. As Andersson and 
T rudgill (1990, 105) explain, “when we talk, we enter into a conversation from our own private 

world of knowledge and beliefs. %e speaker has one such world and the listener another. During 
the conversation, we together build up a shared world of discourse… Small words (fillers) … 
rise like bubbles out of our private worlds to the surface of the shared world. %ey help to give 
indications to do with surprise, irritation, insecurity, and so on, about the relationship between the 
private worlds and the shared one, without forcing us to spell out the whole story word for word”.  
Seen from yet another perspective, fillers may function as discourse markers – markers of how 
the states of information change over the course of a conversation. Discussing presupposes 
exchanging information, “knowledge and meta-knowledge are constantly in flux, as are degrees 
of certainty about, and salience of, information. … What speakers and hearers can reasonably 
expect one another to know, what they can expect about the other’s knowledge of what they 
know, how certain they can expect one another to be about that knowledge, and how salient they 
can expect the other to find that knowledge are all constantly changing” (Schiffrin 1987, 99). 
Fillers give indications of these changes.    
%ese general functions branch out into a number of specific roles that fillers may play, as it will 
be illustrated in what follows.
 As suggested in the previous section, following Schiffrin’s (1987) theory, fillers may sometimes 
function as discourse markers. As such, they contribute to the development of conversation in a 
particular way, by signalling that:
a) the listener has followed the interlocutor’s words and encourages him/her to speak on, 
implicitly acknowledging the temporary dominant role of s/he who has the floor at that point of 
the exchange. In this case, fillers function as backchannels, used, most of the times, though not 
exclusively, “at syntactic and semantic ‘completion points’, where even a takeover would have 
been natural” (Stenström 1994, 6):
(4) LK: Now, Muhammad has gotten progressively worse…
MF: Yes.
LK:… in that he can’t speak – he almost can’t – I don’t think he can speak.
(5) LK: No one ever hit you?
RG: No, no. Which is so interesting because I ended up marrying – I mean 
certainly my mother tried to create a safe, wonderful – or my household was 
wonderful.
LK: Right.
RG: But I married a man like my father…
(6) DT: Well, it seemed that all of the folks went back to Saudi Arabia. %ey 
 got their …
  LK: OK.

DT: %ey put their wives on the planes. %ey all went back to Saudi Arabia.
(7) SG: … this is much more about scope and how widespread this might 
become…and much less about severity.
LK: All right…
SG: So it might not be that serious. It could be mild illnesses.
b) the listener has followed the interlocutor’s words, but, at some point, s/he feels that a complete 
change of topic would be welcome and prompts it with the help of “transition markers” (Stenström 
1994, 21) (thus, implicitly controlling the topical development of the talk): 
(8) DT: It’s a sad thing, but… you have to say to some people “You’re fired. 
%e world is different.” %is is a different world. And … there’s probably no 
business that isn’t laying off people.
LK: Now, you – you write in the book that people refer to you as being 
prescient. And your book also happens to have references to Somali pirates 
who are big news. What do you make of this whole piracy thing? 
(9) Previous topic discussed by the speaker: rehearsing press conferences.
LK (interlocutor): Well, let’s deal with some of the obvious. You’ve got to 
deal with Whitewater, right?
c) the speaker will go on, by introducing a sub-topic to a (general) statement just made or by 
switching to a different topic (connected or not to what has just been discussed). In order to keep 
the floor while s/he hesitates, the speaker uses dummy words in the empty spaces between two 
segments of his/her turn, “much as s/he might drape his/her coat on a seat at the cinema to prevent 
others from taking it” (Deflice 1995, 1), implicitly discouraging the interlocutor’s intervention 
at this point of their dialogue. Should s/he have left a pause unfilled, the interlocutor might have 
interpreted it as a take-over signal, as an indication that the speaker had completed his/her message, 
that there was nothing s/he would have liked to add, and that s/he was ready to give up his/her 
turn. T o prevent being interrupted while planning what to say next, the speaker uses fillers that, 
according to Stenström (1994, 76), “serve as stallers from a discourse-strategic point of view”:   
(10) LK: I’m going to read a little portion… “How we handle difficult situations… 
says a lot about who we are … Some events that will wipe out one person will 
make another person even more tenacious. %at’s why I always ask myself ‘Is 
this a blip or is it a catastrophe?” By the way, … have you ever experienced a 
catastrophe? (sub-topic)
(11) DT: I think unreasonable is the war in Iraq. I mean … hundreds of thousands 
of people killed and maimed and so badly hurt… And, by the way, just to get 
on it, he’s not the one that knocked down the World T rade Center. Saddam 
Hussein did not knock down the World T rade Center. (different topic) 
(12) DT: Here’s a man that not only got elected, I think he’s doing a really good 
job. Now, the sad part is that he can’t just do a good job. He’s got to do a great 
job. Because if he does a good job, that’s not good enough for this country… 
(sub-topic)

(13) LK: … I sat down for an exclusive interview with the former president to talk 
about this earlier today. Now, in the interest of full disclosure, I was a White 
House fellow during his administration. (different topic)
(14) LK: An outstanding group of journalists join us from Washington, Tavis 
Smiley, host of his own show on PBS and best-selling author. His latest 
book is “Accountable: Making America as Good as Its Promise”. Jennifer 
Skalka is editor of “Hotline on Call”. And Chris Cillizza is the White House 
correspondent for “%e Washington Post”… All right, Tavis, we have been 
asking all the guests all day long to give it a – give it a number.   
 Where do you put this, A to D? How’s he doing? (different topic)
d) “the speaker needs to make a new start or rephrase what s/he was going to say in the middle of 
a turn, often because the listener shows that s/he cannot follow or is not convinced” (Stenström 
1994, 131). In such circumstances, “monitor” (Stenström 1994, 131) markers such as I mean 
and actually may help the speaker put things right:
(15)  LK: How do you assess him?
 DT: Well, I really like him. I think that he’s working very hard. He’s  
trying to rebuild  our reputation throughout the world. I mean we really have 
lost a lot of reputation in the world. %e previous administration was a total 
disaster, a total catastrophe. 
(16) DT: … he is really doing a nice job in terms of representation of this country. 
And he  represents such a large part of the country. I mean, to think that a 
black man was going to be elected president – I watched television for years 
where the great political analysts  were saying in 50, maybe in 100 years.
(17)  LK: What’s the reality of that statement, that it must be love?
     RG: Well, the reality is that I – for me…
 LK. Yes.
 RG: … for me is that I didn’t have a model of what a good, healthy  
relationship looked like. I was actually the third generation in my – my family 
experienced domestic violence.
 
e) the speaker is willing to give the floor to the listener, even if for a minimal contribution, by 
indicating that a declarative sentence may as well be interpreted as having the illocutionary force 
of a question that may be answered: 
(18)  DT: We – in fact, I just built a building. You’ve probably seen it*. It’s the
  tallest building in Las Vegas. (*You’ve probably seen it = Have you seen it?) 
   LK. I saw it. 
(19) LK: %e natural scheme of life is temptation. It’s in front of you everyday, 
I suppose*. (*It’s in front of you everyday, I suppose = Is it in front of you 
everyday?)
  JO: Sure.

f) the speaker is willing to give the floor to the listener, for the latter to confirm or invalidate the 
former’s opinion, as the case may be. %e type of fillers that fulfil this function is represented by 
tag questions (whose intonation one may guess on the basis of the interlocutor’s response, in the 
absence of indications of intonation in the transcripts) and adverbs with the same function as tags:
(20) MF: I was diagnosed in ’91 and it wasn’t until ’98 that I admitted publicly 
that it was a situation I was facing. 
  LK: It is not life-threatening, is it?
  MF: No.    
(21) LK: But you’re not fire and brimstone, right? You’re not pound the decks and 
hell and damnation?
  JO: No. %at’s not me. It’s never been me. I’ve always been an encourager at 
heart…
  I know there is condemnation but I don’t feel that’s my place.
  LK: You’ve been criticized for that, haven’t you? 
  JO: I have. I have. Because I don’t know.
  ...
  LK: Good news guy, right?
  JO: Yeah. But you know what? It’s just in me. I search my heart and I think, 
God, is this what I’m supposed to do?  
 When they express attitudes on the part of the speaker, fillers may be said to function as interactional 
signals, in as much as the expression of particular emotions contributes to the carrying on of 
the conversation along certain psychological mood lines. %e linguistic manifestation of the 
following feelings may be illustrated with examples from the corpus analyzed: 
a) the speaker’s confidence, strong belief in his/her own opinion, introduced and/or rounded off 
by the filler (the speaker implicitly acts as authoritative and signals that s/he wants his/her point 
of view to be accepted by the interlocutor):
(22) RG: I’m telling you, you cannot raise 14 children as a single parent.
TS:… somebody asked me earlier tonight what I thought the story line out 
of this was. I have already seen some storylines. Clearly, the comment about 
Chrysler … that is going to be news tomorrow, for sure.  
(23) LK: Do you assess him as a champion?
DT: Oh, yes. He’s a champion… He’s totally a champion.
 (24) LK: What does he have?
  MD: He obviously has a great intellect. And, man, he has a terrific capacity  
  to persuade. 
 (25) LK: But you’re not going to march with the tea party?
  DT: I don’t march with the tea party. But I’ll tell you what, they have a good 
point, because when you see this kind of money… I can understand where 

they’re coming from. 
 (26) LK:  Always – do you remain – it is easy to be an optimist?
  MF: … for me, it’s second nature. It’s just the way I look at life. And I – it’s 
certainly a challenge now for most people to be optimistic, obviously, with all of the troubles 
we have and the problems that the country is facing.
b) the speaker’s belief that his/her words are important and should be considered so by the 
interlocutor (the speaker explicitly or implicitly aims at drawing the interlocutor’s attention on 
what s/he is saying; fillers have a “focus” function in this case, as Fleishman and Yaguello (2004) 
suggest, i.e., they mark the information coming to their right as focal):
(27) LK: … do you see any chance of the Mideast having peace?
 BC: I do…
 LK: George Mitchell has a lot to do with it?
 BC: Well, he can. He’s going to have to fill in the blanks, you know, with 
people. But, look, however the Israeli government is constituted… there are 
two things that give us hope.
(28) PA: #e point here is that this is an era of – a political era, but also a media era, 
because Saddam and his government were very media savvy.
(29) LK: What do you make of Sarah Palin, your running mate, apparently making 
– not apparently – making critical statements about the campaign?
 McC: Listen … I’m honoured that she would run with me. And there’s – look, 
whenever there’s a losing campaign, there’s always a little bit of back and forth 
that happens post mortem. Look, I’m so grateful to have her as a friend…  
(30) LK: Denise, Nicole’s younger sister, has written a Web exclusive. So go to CNN.
com … to read TB’s relevant comments. Oh, and if you have a personal story 
you would like us to share with our audience, that’s the place to tell us.
c) hesitation, lack of full confidence on the part of the speaker or intention to make a hedged 
statement. A possible pragmatic implication here is that the speaker does not want to take full 
responsibility for the statement made or that s/he considers the information conveyed to have 
too heavy an emotional load to be delivered directly. On the other hand, hesitation may be 
caused by the fact that no “option offered by the question provides a sufficient basis from which 
to choose an answer” (Schiffrin 1987, 106). Among the reasons that might cause respondents 
to diverge from the choices offered them by a prior question, there might be a lack of adequate 
information or knowledge, an inaccurate assumption by the questioner, a complication of the 
situation being questioned (the examples below are illustrative of all these situations):
(31) LK: A headline, Dr. Sanjoy, in the “L.A. Times” today said that is not as bad 
as they’re saying. Is that right? 
SG: Well, you know, it’s hard to say. You know, one thing about the news 
yesterday, Larry – you and I talked about this idea that we are having an 
imminent pandemic. You know, it’s frightening to hear that, I think, for 
anybody.  What I think is important to sort of remind people is what 

they’re really referring to, I think after talking to folks at the World Health 
Organization, is that this is much more about scope… 
(32) LK: Robin, do you buy the comparison to Jackie Kennedy?
RG: Well, in some ways, I think that it’s fair to compare them, only because 
it’s a way for us to measure the different ways in which women in public life 
have progressed…
(33) LK: How would you describe this White House? Is she right-confused?
WB: Well, I think that there are several issues right now that are the most  
 important.
(34) LK: Now, let’s dig into it, Rita. Up and down, polls up, polls down – how 
would you describe this White House?
RB: Well, I…
LK: Is it beleaguered?
RB: Well, right now, that they are in a mode of reassessment.
(35) LK: What should his punishment be?
 BS: Well, I mean. Giving my personal opinion. I remember the Nuremburg 
trials after the war, and I was in favor. But I think pinning it all on one 
person … is not as good as Archbishop T utu, who talks about truth and 
reconciliation.  
Unlike in the examples above, there are cases when fillers are not used to express the speaker’s 
attitude towards his/her own statements, but rather towards what the interlocutor has just said. 
Such an attitude may be:
a) surprise, as a result of having received unanticipated information:
(36) CS: OK, I T witter.
LK: You T witter? You do? You’re so modern. And you have a blog?
CS: Yes.
LK: My gosh, where do you find the time?
(37) RG: … when I wrote my book, I was somewhere talking about something. 
And a woman came up to me and she said I wanted you to stop talking 
because I felt like everybody would know that you were talking about me. 
LK: Oh, boy. It seems – and it’s the same progression, too – it seems. 
b) disagreement: 
(38) LK: You just seemed a little hyperactive.
MF: Oh. What you’re seeing here is simply my natural boyish exuberance,
that’s all. 
(39) LK: %ey say you have regrets about endorsing McCain.
DT: Come on, I’ve known John for a long time. And I’ve never met our 
current president. And I look forward to doing it. But I’ve known John, 
really, for a long time. He’s a good, he’s a quality man. So, certainly, I have 
no regrets about that.  

c) annoyance: 
(40) LK: … but you had the opportunity to speak to your client these last days.
MD: For god’s sake, we have been denied that opportunity as we’ve been 
denied almost every other right of the defendant throughout this trial.  
(41) LK: … will his [Saddam’s] execution affect the popularity of President Bush?
MW: Within Iraq or within this country?
LK: Within this country.
          MW: I guess not.
          LK: Maybe in both countries.
MW: Oh, dear, no. As I said earlier, we are so far beyond the act of toppling 
Saddam. … I doubt it very much that it will bring back popularity for the 
president.
%ough impossible to classify as expressions of stance either towards the speaker’s expressed 
view or towards the content of the interlocutor’s intervention, fillers such as you know and 
you see may be mentioned as illustrative of another type of attitude on the part of their user, 
directed, this time, towards the interlocutor himself/herself – the desire to intensify the 
former’s relationship with the latter. %e socializing strategy that involves the use of such fillers 
contributes to creating a climate favourable for a smooth continuation of the interaction. 
%erefore, you know and you see thus used may be considered interactional signals. Quotes (42) 
and (43) below exemplify this usage:
(42) LK: You know, it’s the cycle of domestic violence, which is about the power 
and control of one human being over another – the verbal, the emotional, the 
psychological abuse, you know, the chipping away at one’s self-esteem – you’re 
stupid, you’re ugly, you’re worthless, you’re no good. I’m sure you heard it 
Robin, you know, nobody is going to want you…  
(43)  DA: But I have to say, you see, when we started the campaign for president, 
it wasn’t an easy transition for him. %ere were months of adjustment… 
I haven’t seen that this time… He’s comfortable. He’s in command. And, 
you know¸ there isn’t a day that goes by when I don’t … think I’m glad he’s 
there. 
%ere are other fillers, too, which, though not intended for strengthening the relationship 
between the participants in conversation, are, however, linguistic embodiments of emotions that 
replace spelled out feelings of surprise, excitement, admiration (in my data, manifested towards 
things that appeared in videos shown during the interviews and, therefore, parts of some side 
comments). %ese are also beneficial to the efficient flow of the dialogue. Examples (44) and (45) 
are illustrative in this sense:
 (44) LK: Have you ever been to the White House?
  CS: Yes.
  LK: What did you think of it? Oh, my gosh. A lemon tree. Lemons are my  
  favorite. %at’s what I drink my vodka with. 

 (45) LK: So this is the very infamous gift wrapping room. Oh, my goodness.  
  Do you actually wrap the gifts yourself?  
  CS: Oh, yes.  
 As explained and exemplified above, though fillers may not carry a heavy load of meaning, it 
cannot be denied that they are often functional in conversation. At least two main roles of fillers 
have been identified in the ten Larry King interviews that I concentrated on in the present 
paper. 
On the one hand, they function as discourse management elements, or discourse markers, in 
the sense that their particular use influences the development of the topic and the succession of 
turns. By using fillers, the listener may support the speaker’s going on or, on the contrary, s/he 
may prompt a complete change of topic; the speaker may hint at his/her intention of talking on, 
either by introducing a sub-topic of what is being discussed or by switching to a new topic, at 
his/her intention to start over or to reword what s/he was going to say or at his/her willingness to 
yield the floor to the interlocutor, if s/he is ready to take it. 
On the other hand, they are signals of the speaker’s various emotional attitudes (either towards the 
content of his/her own words, towards what has just been said or towards things shown during 
the interview), which may contribute to pushing forward the discussion against a favourable 
psychological background, or to supporting the intention to strengthen the relationship between 
the speaker and his/her interlocutor, which may have the same effect. 
%e examples offered by the interviews scrutinized also prove that the same filler may be 
multifunctional. It may fulfil different functions, depending on the context in which it is used – 
the adverb now, for example, may function both as a way of the speaker’s signalling that a change 
of topic is prompted, as in (8), or as a marker of his/her intention to go on with sub-topics of a 
main issue just introduced, as in (11) and (12), while the adverb right may be indicative, on the 
one hand, of the fact that the listener has followed the interlocutor’s words and encourages him/
her to go on, as in (5) and of his/her willingness to yield the floor for validation or invalidation 
of his/her opinion by the interlocutor, on the other, as in (21).  
%is makes one wonder whether the term filler might not be used as an umbrella that covers in 
fact three main uses of such words that have no semantic-referential meaning: that of filling a 
conversation gap, that of discourse marker responsible for the flow of conversation, and that of 
marker of emotional attitudes.